Classification of skin sensitizing substances: a comparison between approaches used by the DFG-MAK Commission and the European Union legislation.

A systematic classification of substances (or mixtures of substances) with regard to various toxicological endpoints is a prerequisite for the implementation of occupational safety strategies. As its principal task the "Commission for the Investigation of Health Hazards of Chemical Compounds in the Work Area" of the "Deutsche Forschungsgemeinschaft" (DFG-MAK Commission) derives and recommends maximum workplace concentrations and biological tolerance values (MAK and BAT values) based exclusively on scientific arguments. Several endpoints are evaluated separately in detail, e.g. carcinogenicity, risks during pregnancy, germ cell mutagenicity or contribution to systemic toxicity after cutaneous absorption. Skin- and airway sensitization is also considered; the present paper focuses on these two endpoints.

A 2 year oral carcinogenicity study with cadmium on rats.

A longterm 2-year feeding study on cadmium using CdCl2 . 1H2O has been performed on Wistar-rats. The tested Cd2+-dietary levels were 1, 3, 10 and 50 ppm, respectively. Fifty male and 50 female rats were used for each of the levels. Hundred rats of each sex served as controls. No effects occurred on food intake and survival rate at any dietary level. Growth was unchanged up to 10 ppm whereas 50 ppm resulted in lower weight gain in males only. Histopathological evaluation of a large variety of tissues revealed a number of tumours. Cadmium administered orally was not associated with an increased incidence of total numbers of tumours or of any specific type of neoplasia, although the highest level tested resulted in adverse effects.

Chronic effects on single and multiple oral and subcutaneous cadmium administrations on the testes of Wistar rats.

Thirty mature male Wistar rats were administered cadmium as CdCl2 X 1H2O in single doses of 50 mg/kg p.o. or 2.5 mg/kg i.p. or 10 weekly doses of 5 mg/kg p.o. or 0.25 mg/kg i.p., respectively. Ten males, each treated correspondingly with the vehicle, served as control groups. Some of the animals were necropsied after 12 and 18 months, respectively, the remainder were kept for up to 30 months. In a supplementary study 25 males were each treated once with 200 mg/kg p.o. or 2 mg/kg s.c. and 35 males with 100 mg/kg p.o. This experiment was terminated after 6 months. Animals having received 1 X 100 or 1 X 200 mg/kg p.o. or 1 X 2.0 or 1 X 2.5 mg/kg s.c. showed severe lesions of the whole testicular parenchyma with massive calcification of the necrotic tubuli and pronounced fibrosis of the interstitium. All animals receiving 2.5 mg/kg s.c. had a Leydig cell tumor in at least one of the testes. In 5 out of 15 animals surviving 18 months these tumors were classified as malignant (mean time of induction: 858 +/- 77 days). All the other tumors detected were not regarded as causally related to treatment. The results of the supplementary study indicate that with very high oral cadmium dosages Leydig cell tumors may be induceable. Since the massive lesion of the testes appears to be the prerequisite for the occurrence of induced Leydig cell tumors, a non-genetic mechanism is to be assumed in respect to the formation of these tumors.

Thirteen-week subchronic rat inhalation toxicity study with a recovery phase of trivalent chromium compounds, chromic oxide, and basic chromium sulfate.

The toxicity of trivalent chromium compounds; chromic oxide and basic chromium sulfate, was investigated in rats in a 13-week nose-only inhalation study that included a 13-week recovery period. Nose-only exposures to insoluble chromic oxide dust at 4.4, 15, or 44 mg/m3 or soluble basic chromium sulfate dust at 17, 54, or 168 mg/m3 (trivalent chromium equivalent concentrations of 3, 10, and 30 mg/m3) were carried out for 6 h/day, 5 days/week. No compound-related mortality occurred. General toxic effects, only observed with high-exposure levels of basic chromium sulfate, included sporadic signs of labored breathing and depressed body weights. No apparent compound-related effects were noted for sperm motility or morphology, for any concentration of either test material. Bronchoalveolar lavage fluid evaluations showed test material in mononuclear cells with chromic oxide and increased neutrophils, protein, lactic dehydrogenase and cellular debris with basic chromium sulfate. The principle effects for both materials were primarily to the respiratory tract. Chromic oxide caused pathological changes in the bronchial and mediastinal lymphatic tissue and lungs, consisting of the presence of pigment-laden macrophages, lymphoid and septal hyperplasia, and interstitial inflammation similar to that observed with other inert dusts. Basic chromium sulfate produced more severe and widespread effects in the nasal cavity, larynx, lungs, and mediastinal lymph node. Effects were characterized by accumulation of foreign material, infiltration of alveolar macrophages, septal cell hyperplasia, and granulomatous and chronic inflammation. Pigment was still present in chromic oxide and, to a lesser extent, in basic chromium sulfate-treated animals after the 13-week recovery period, with partial recovery of the pathological lesions. A NOAEL was not established for either test material, but 4.4 mg/m3 was thought to be near the NOAEL level for subchronic exposure to chromic oxide. The results of this study indicate significant differences in toxicity to the respiratory tract between trivalent chromium compounds chromic oxide and basic chromium sulfate. These are likely related to differences in acidity and water solubility, rather than chromium concentration per se. This conclusion is substantiated by the lack of effect on other internal organs.

Time course of enzyme induction in liver and kidneys and absorption, distribution and elimination of 1,4-dichlorobenzene in rats.

Time course of enzyme induction was measured in Fischer344 rats treated daily at 150 and 600 mg 1,4-dichlorobenzene (1.4-DCB)/kg p.o. up to 28 days. The monoxygenases 7-ethoxycoumarin O-deethylase (ECOD), 7-ethoxyresorufin O-deethylase (EROD) and aldrin epoxidase (ALD) as well as the phase II enzymes; epoxide hydrolase (EH), glutathione S-transferase (GS-T) and glucuronyl transferase (GLU-T) were dose-dependently induced in the liver of males and females. A pronounced induction in the kidneys was measured at 600 mg/kg only for ECOD. After single oral administration of 100 and 1000 mg/kg bw and feeding of 100 and 1000 ppm (corresponding to approximately 10 and 100 mg/kg bw) to male Wistar rats for 28 days, the time course of 1,4-DCB and 2,5-DCP concentrations was investigated in plasma, adipose, hepatic and renal tissue. In addition, total urinary excretion of 2,5-DCP was determined. After single application, 1,4-DCB and 2,5-DCP were rapidly eliminated from the plasma and tissues, 40-60% of the dose administered was excreted as 2,5-DCP in the urine. There were no indications of cumulative effects after a feeding period of 28 days. The concentrations decreased in all tissues until the 7th day of study. Thereafter, there seems to be a steady state until the 28th day. A total of 7 days after the end of exposure, no more residues could be detected. Following long-term inhalation (450 and 3000 mg/m3) 1,4-DCB concentrations were highest in adipose tissues at 6 months followed by a marked decline at 18 months. 1,4-DCB and 2,5-DCP concentrations in plasma and liver were much lower but again with a peak at 6 months. When compared with published human data on measurements in plasma, urine, liver and adipose tissue the results suggest that there should be no hazard for the general population.

Toxicity of 2,3,7,8-tetrabromodibenzo-p-dioxin in rats after single oral administration.

Five male and female rats per dose-group received 2,3,7,8-tetrabromodibenzo-p-dioxin (2,3,7,8-TBDD) once on the first day of the study. Doses of 10, 33, 100, or 300 micrograms 2,3,7,8-TBDD/kg body wt. and the vehicle control were administered by gavage. About 20% of 2,3,7,8-TBDD was excreted via feces. Severe body weight retardation was observed in the 100 and 300 micrograms/kg dose-groups. Most animals in the 300 micrograms/kg dose-group and the females receiving 100 micrograms/kg showed emaciation, rough coat and a poor health (wasting syndrome). Of the animals dosed with 300 micrograms/kg, 3 males and all females died. After 100 micrograms 2,3,7,8-TBDD/kg 3 females died. Measured 4 weeks after dosing, triiodothyronine (T3) was increased and thyroxin (T4) was reduced dose dependently in serum. A dose-dependent decrease in thymus weights was observed at necropsy and histological examinations showed that thymus and spleen were depleted of mature lymphocytes. An increase in liver-to-body weight ratio was observed in all dose-groups. The histological examination revealed hypertrophy of centrilobular hepatocytes in the liver of animals treated with 100 micrograms/kg, which was less severe at the 33 micrograms/kg dose. Hypertrophic hepatocytes were also detected in some animals at the lowest dose. Induction of enzyme activities of the mixed function oxidases ethoxycoumarin O-deethylase (ECOD), ethoxyresorufin O-deethylase (EROD) and aryl hydrocarbon hydroxylase (AHH) in liver tissue differed for each of the three enzymes. Two days after administration, enzyme activities were increased but did not differ substantially between dose-groups. Twenty-eight days after dosing the increase in activity after 10 micrograms/kg was largest and the EROD of the 100 micrograms/kg dose-group in females was close to that of the control. This inverse dose-response relationship may be due to impaired liver cell function at higher doses.

Subchronic toxicity of 2,3,7,8-tetrabromodibenzo-p-dioxin in rats.

2,3,7,8-Tetrabromodibenzo-p-dioxin (2,3,7,8-TBDD) was administered daily to male and female rats for 91 days by gavage. Ten male and 10 female rats per group received 0.01, 0.1, 1, 3, or 10 micrograms 2,3,7,8-TBDD/kg body weight per dose per day, solubilised in arachis oil. At 1 microgram/kg per day and above, body weight gain was dose-dependently reduced by treatment. Animals in the 3 and 10 micrograms/kg dose groups showed symptoms of wasting syndrome. Fifty percent of the animals in the 3 micrograms/kg dose-group died and all animals of the highest dose (10 micrograms/kg) died or had to be killed in extremis. Hematological investigations indicated changes--mainly in the 1 and 3 micrograms/kg dose-groups--in hemoglobin content, packed cell volume and number of thrombocytes. The prothrombin-time was markedly prolonged after 3 micrograms/kg in week 13. Clinical chemistry performed at the end of treatment revealed an increase in plasma alkaline phosphatase (APh), aspartate aminotransferase, ASAT and alanine aminotransferase, ALAT (females only) in the highest surviving dose-group (3 micrograms/kg). Marginal changes of APh and ASAT were seen in rats in the 1 microgram/kg dose-group. In the same animals, total bilirubin was elevated. Triglycerides were reduced mainly at 1 and 3 micrograms/kg. Serum thyroxin was reduced, beginning with a marginal change at 0.1 micrograms/kg, triiodothyronine was elevated, starting with a dose of 1 microgram/kg. Thymus weights were reduced in rats of the 1, 3 and 10 micrograms/kg dose-groups. Histopathological analysis showed atrophy of the lymphatic tissue in thymus and spleen. Investigations of the liver indicated peliosis hepatis after treatment with 3 or 10 micrograms/kg. Activities of microsomal enzymes (ethoxyresorufin O-deethylase, ethoxycoumarin O-deethylase, aryl hydrocarbon hydroxylase, UDP-glucuronyltransferase) investigated in liver, lung and kidney were dose-dependently elevated after 13 weeks of treatment. At a dose of 3.0 micrograms/kg, activities were below those of the dose 1.0 microgram/kg, probably due to liver toxicity. The induction ratio of kidney was generally higher than in liver and lung. No signs of treatment-related toxicity were observed in the 0.01 and 0.1 micrograms/kg groups after the subchronic administration of 2,3,7,8-TBDD by gavage.

Subchronic oral toxicity and analytical studies on nickel rutile yellow and chrome rutile yellow with rats.

The inorganic pigments nickel rutile yellow and chrome rutile yellow were fed to groups of 15 male and 15 female rats each for 3 months at dietary levels of 0, 10, 100, 1000 and 10000 ppm, respectively. On an additional 5 animals each, levels of nickel and antimony or chromium and antimony in liver and kidneys, respectively, were measured after 1 and 2 months. Appearance, behaviour, food consumption, growth, mortality, haematological and clinical chemical data, organ weights, and gross and micromorphology of organs were not affected in any dose group. In livers and kidneys antimony median levels below 30 ppb were detectable only in the group of rats fed the highest level of 10000 ppm of the two pigments.

Damage to mitochondrial DNA induced by the hepatocarcinogen diethylnitrosamine in ovo.

Different doses of the hepatocarcinogen diethylnitrosamine (DEN) were injected into fertilized turkey eggs 8 days before hatching. The embryos were removed from the eggs after 4 days and liver samples were shock frozen. Mitochondrial DNA (mtDNA) was purified from the samples. Electrophoresis on agarose gels with native mitochondrial DNA and with ribonuclease-treated mitochondrial DNA revealed a DEN-induced effect on the molecular size of the mtDNA. The content of mtDNA of the regular size of 16 kb dose-dependently decreased, whereas the amount of mtDNA fragments of various size increased. Fluorescent staining of the electrophoresis gels allowed the densitometric quantification of the mitochondrial DNA of the regular band at 16 kb and the amount of fragments of irregular size (smear). The diethylnitrosamine-induced effect was dose-dependent over the whole dose range from 1.24 to 6.2 mmol/kg. Even the lowest dose (10 mg DEN per egg) showed clear-cut effects. Mitochondrial damage and malfunction may be mechanistically involved in the neoplastic transformation and in aging phenomena. The in ovo model is a simple and rapid approach for investigations on chemically induced alterations of mtDNA.

[Subchronic toxicologic studies using cyclamate-Na in dogs]. / Subchroische toxikologische Untersuchungen mit Cyclamat-Na an Hunden

2 male and 2 female beagle dogs were fed cyclamate-Na in daily doses of 0 (control), 150, 500, 1500 mg/kg of body weight with the diet. Liquid feces occurred only after feeding 1500 mg/kg/day during the first few days of the experiment. Haematological, clotting and clinical chemistry parameters were unchanged in the groups up to the level of 1500 mg/kg/day. No effects were noted in carbohydrate and fat metabolism, nor in ECG or blood pressure. Gross pathology and histopathology did not reveal any change in any dose group. Daily fed doses of cyclamate-Na up to 1500 mg/kg body weight were well tolerated for three month in the dog.

Study by fluorescence microscopy of the effect of fluorescent whitening agents on the skin of mice.

Fluorescence microscopic studies of the skin of hairless mice showed that a fluorescent whitening agent (FWA) of the bis(phenyltriazolyl)stilbenedisulfonate type did not penetrate into the subepithelial layers (dermis and subcutaneous tissue) of the skin after cutaneous application.

Foci of altered hepatocytes induced in embryonal turkey liver.

Fertilized turkey eggs were injected with N-nitrosomopholine (1-10 mg/egg) or urethane (0.1-10 mg/egg) prior to or during the first hours of incubation. Four days prior to hatching, the experiments were terminated and the livers of the embryos removed. The hepatocarcinogens induced focally arranged hepatocytes with morphological and metabolic changes that were similar to preneoplastic foci observed in rodent studies in vivo. Foci with enlarged cells and excessive storage of glycogen as well as basophilic foci with small cells, low in glycogen content and with frequent mitotic figures were found. The in ovo transformation of embryonal hepatocytes may be a promising rapid model of hepatocarcinogenesis.

Dose dependence of diethylnitrosamine-induced nuclear enlargement in embryonal turkey liver.

Avian embryos (turkey) were exposed to diethylnitrosamine in ovo. On the first day of incubation doses of 0.5-5.0 mg/egg were injected into the white of the fertilized egg. The experiment was terminated 4 days before hatching. Livers were removed and prepared for subsequent histological examination. In haematoxylin and eosin stained sections the areas of hepatocyte nuclear profiles were measured by semi-automatic image analysis. In liver samples of diethylnitrosamine-exposed embryos hepatocyte nuclei of more than twice the size of normal hepatocyte nuclei were found. The incidence of the enlarged nuclei was clearly dose dependent. An increase in the size of hepatocyte nuclei was observed after low doses of diethylnitrosamine that did not induce common signs of non-specific toxic effects, e.g. cell death, fat vacuoles or loss of glycogen. The slope of the dose-response curve was rather steep. A 10-fold increase in the dose of the carcinogen resulted in a 100-fold increase in the incidence of enlarged hepatocyte nuclei. In combination with preneoplastic foci of altered hepatocytes, the quantification of nuclear enlargement can provide a valuable complementary parameter for the evaluation of carcinogen-induced effects in ovo.

Glycogen phosphorylase hyperactive foci of altered hepatocytes in aged rats.

In untreated 12- to 24-month-old rats, the enzyme histochemical pattern of 45 focal hepatic lesions was investigated in serial sections. In addition to previously characterized glycogen storage foci, a new type of enzymatically altered hepatic focus was found. The outstanding feature of this was an increased glycogen phosphorylase activity. The frequent appearance of glycogen phosphorylase hyperactive foci simultaneously exhibiting excessive glycogen storage suggests a close relationship to the well known glycogen storage foci representing an early stage in the sequence of cellular changes which lead to hepatic tumors.

Nonlinearity of nuclear enlargement in hepatocytes induced by the carcinogen N'-nitrosomorpholine in Ovo.

In this study we investigated the effects of different doses of the carcinogenic nitrosamine N'-nitrosomorpholine (NNM) on the occurrence of enlarged nuclei in embryonic turkey liver in order to evaluate whether this parameter might represent a quantitative indicator of chemically induced hepatocarcinogenesis. Therefore fertile embryo turkey eggs were injected with NNM over a dose range of 125 microg-8 mg/egg at the first day of incubation. After incubation for 24 days, the embryonic livers were removed and processed for histologic evaluation. The induction of hepatocytes with enlarged nuclei (nuclear profiles > 35 microm2 was quantitated morphometrically in hematoxylin and eosin (H&E)-stained sections. The NNM treatment increased both the number of enlarged hepatocyte nuclei and the areas of the individual profiles of the enlarged nuclei in a dose-dependent manner. Exposure to 500 microg-8 mg NNM/egg resulted in a statistically significant increase in the number of hepatocytes with enlarged nuclei. The lower doses of 250 microg and 125 microg NNM/egg showed a similar albeit not significant trend. Signs for cytotoxic effects on the hepatocytes, such as necrosis or enhanced cytoplasmic vacuolization, were observed in tissue samples of embryos exposed to 4 or 8 mg NNM, but not after treatment with lower doses. The dose-effect curve for the induction of the nuclear enlargement was nonlinear, with a moderate slope for lower dose levels of 125-500 microg/egg and a steep slope for higher dose levels of 1-8 mg. Findings in rodents indicate a pathogenic link between the occurrence of enlarged nuclei and hepatocarcinogenesis. Based on the results with NNM, it is suggested that the in ovo model may represent a rapid, convenient, and inexpensive experimental approach for dose effect investigations on chemically induced hepatocarcinogenesis.

Acute, subacute, and subchronic inhalation toxicity studies of respirable polymeric methylene diphenyl diisocyanate (polymeric MDI) aerosol in rats.

Short-term inhalation toxicity studies with respirable polymeric methylene diphenyl diisocyanate (polymeric MDI) aerosol were performed in rats. The 4-hr LC50 was found to be 490 mg polymeric MDI/m3 (95.5% < 4.3 microns). Exposure of (4-week-old) rats to 0, 2.2, 4.9, or 13.6 mg polymeric MDI/m3 (95% < 5 microns) for 2 weeks resulted in mortality, severe growth retardation, and elevated lung weights at 13.6 mg/m3; at 4.9 mg/m3 slight growth retardation and slightly elevated lung weights were observed. A 13-week study with 6-week-old rats exposed to 0.35, 1.4, or 7.2 mg polymeric MDI/m3 (95% < 5 microns) revealed transient growth retardation and a slightly increased number of pulmonary alveolar macrophages occasionally accompanied by increased numbers of mononuclear cells and fibroblasts in alveolar septa only at 7.2 mg/m3. In a second 2-week study with 4- or 6-week-old rats exposed to 14.1 mg polymeric MDI/m3 (95% < 5 microns), 4-week-old rats died earlier and in greater numbers than 6-week-old rats. In a second 13-week study with 6-week-old rats, using exposure concentrations of 0, 4.1, 8.4, and 12.3 mg polymeric MDI/m3 (95% < 5 microns) and including a 4-week recovery period, 12.3 mg/m3 induced mortality, growth retardation, severe respiratory distress, increased lung weights, degeneration and hyperplasia of the nasal epithelium, accumulations of macrophages in the lungs and mediastinal lymph nodes, and focal inflammatory changes in the lungs.(ABSTRACT TRUNCATED AT 250 WORDS)

Effects of low doses of N-nitrosomorpholine on the development of early stages of hepatocarcinogenesis.

Male Sprague-Dawley rats received the hepatocarcinogen N-nitrosomorpholine (NNM) in the drinking water at low dose levels ranging from 6 mg/l to 60 mg/l for 6 and 12 weeks, respectively. Foci of altered hepatocytes (FAH) were demonstrated histochemically using changes in the activities of glucose-6-phosphate dehydrogenase and glycogen phosphorylase, and in the glycogen content as markers. Proliferating cells were detected by the immunohistochemical reaction for proliferating cell nuclear antigen (PCNA). The number and size of foci of altered hepatocytes increased in a time and dose-related manner. The dose-effect curves were non-linear with a slight positive slope at the low doses and a markedly increased slope at higher doses. The number of PCNA positive hepatocytes showed a dose-dependent increase. In addition to the granular distribution of PCNA in the nuclei, hepatocyte nuclei with homogeneously distributed PCNA occurred in animals exposed to 60 mg/l NNM. It is proposed that these cells are related to the occurrence of hepatocytes with higher ploidy induced by NNM and may be regarded as cells in the G2 phase of the cell cycle. The non-linear shape of the dose-response-curve of the FAH suggests that some mechanisms contribute to carcinogenesis over the whole dose range, whereas other mechanisms enhance carcinogenesis only at higher doses. The relevance of the non-linear dose-effect curve for the risk assessment of carcinogens is discussed.

Rapid induction of preneoplastic liver foci in embryonal turkey liver by diethylnitrosamine.

Diethylnitrosamine (DEN) was injected into fertilized turkey eggs. After incubation for 24 days, the livers of the embryos were removed and investigated by means of histological and enzyme-histochemical methods. Doses of 2 and 5 mg DEN per egg had no effect on mortality and embryo weight but induced various types of foci of altered hepatocytes. In addition, a replacement of the trabecular structure of the liver by a tubular arrangement of the hepatocytes and markedly enlarged hepatocytes were found. The diethylnitrosamine-induced foci of altered hepatocytes were very similar to the preneoplastic lesions that occur in the liver of mammals during hepatocarcinogenesis and are regarded as early indicators of carcinogenicity. The presented in ovo model is a simple (1 dose), rapid (24 days), and inexpensive (no animal housing) approach for the induction of foci of altered hepatocytes. The sensitivity of this carcinogenesis bioassay appears to be comparable to that of lifetime studies in rodents. It is proposed that the induction of preneoplastic foci of altered hepatocytes in ovo is a valuable screening model for hepatocarcinogenic effects.

IL-10 down-regulates T cell activation by antigen-presenting liver sinusoidal endothelial cells through decreased antigen uptake via the mannose receptor and lowered surface expression of accessory molecules.

Our study demonstrates that antigen-presenting liver sinusoidal endothelial cells (LSEC) induce production of interferon-gamma (IFN-gamma) from cloned Th1 CD4+ T cells. We show that LSEC used the mannose receptor for antigen uptake, which further strengthened the role of LSEC as antigen-presenting cell (APC) population in the liver. The ability of LSEC to activate cloned CD4+ T cells antigen-specifically was down-regulated by exogenous prostaglandin E2 (PGE2) and by IL-10. We identify two separate mechanisms by which IL-10 down-regulated T cell activation through LSEC. IL-10 decreased the constitutive surface expression of MHC class II as well as of the accessory molecules CD80 and CD86 on LSEC. Furthermore, IL-10 diminished mannose receptor activity in LSEC. Decreased antigen uptake via the mannose receptor and decreased expression of accessory molecules may explain the down-regulation of T cell activation through IL-10. Importantly, the expression of low numbers of antigen on MHC II in the absence of accessory signals on LSEC may lead to induction of anergy in T cells. Because PGE2 and IL-10 are released from LSEC or Kupffer cells (KC) in response to those concentrations of endotoxin found physiologically in portal venous blood, it is possible that the continuous presence of these mediators and their negative effect on the local APC may explain the inability of the liver to induce T cell activation and to clear chronic infections. Our results support the notion that antigen presentation by LSEC in the hepatic microenvironment contributes to the observed inability to mount an effective cell-mediated immune response in the liver.

Regulation of endotoxin-induced IL-6 production in liver sinusoidal endothelial cells and Kupffer cells by IL-10.

Sinusoidal endothelial cells and Kupffer cells are the first cell populations in the liver that come into contact with gut-derived endotoxin in portal blood. Although endotoxin concentrations as high as 1 ng/ml are physiologically present in portal blood, no local inflammation is seen. We show that the proinflammatory cytokine IL-6, which is central to the development of inflammatory reactions in the liver, is produced by sinusoidal endothelial cells and Kupffer cells in response to low concentrations of endotoxin (100 pg/ml to 1 ng/ml). The anti-inflammatory cytokine IL-10 down-regulated endotoxin-induced IL-6 release in endothelial and Kupffer cells. Importantly, Kupffer cells secreted IL-10 after endotoxin stimulation and may therefore participate in the local regulation of inflammation. We have found that IL-6 secretion in Kupffer cells is tightly regulated by endogenous IL-10, because increased IL-6 secretion resulted when neutralizing antibodies to IL- 10 were added to resting and endotoxin-challenged Kupffer cells. Furthermore, repeated exposure of endothelial cells to endotoxin induced a state of tolerance which resulted in decreased release of IL-6 in response to a second endotoxin challenge. Our results support the notion that inflammatory reactions in the liver in response to endotoxin are down-regulated by local release of the anti-inflammatory cytokine IL-10 that is produced by Kupffer cells.